The support system for making medical decisions is a dynamically developing knowledge base at the intersection of medical sciences, medical institutions, doctors, etc. The knowledge base should be replenished, as clinical cases, recommendations, and treatment algorithms are being improved. In addition, in the current situation there is a clear trend towards personalized medicine, which will be the most productive. The article discusses the aspects of using medical decision support systems in the practice of medical organizations. The necessity of using the Unified State Health Information System is emphasized. The aspects of standardization of medical information systems and the use of ontologies, which are open standards for the description of the semantic components of the treatment process, are noted. LOD technology specifications are given. The possibility of combining distributed semantic repositories created by various organizations (health authorities, medical organizations, research centers, etc.) into a single network is noted. The possibility of using UMLS for the development of information retrieval systems, for registration of patients, to facilitate communication between various systems that analyze biomedical information is being considered. The task of the near future is to develop a medical information system as an intellectual physician assistant, including the introduction of clinical guidelines, automatic quality control of medical care and other areas that can significantly improve the quality of medical care and reduce the number of medical errors. Three groups of criteria for the quality of the doctor's activity are defined: semantic, temporary and effective. It is noted that the widespread use and constant replenishment of modern clinical knowledge is based primarily on clinical recommendations. The experience of using the register for stroke is presented and the results of clinical recommendations tested in practical public health in the field of neurology are taken into account.